[
https://issues.apache.org/jira/browse/OOZIE-2817?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15939029#comment-15939029
]
Peter Cseh commented on OOZIE-2817:
-----------------------------------
Thanks [~abhishekbafna], for the investigation!
You're right about FAILURES vs failures.... I'll fix it with an other amend
commit I guess :)
I don't know why the number of tests is different in Jenkins and in our report.
I haven't touched that part but it looks like it just sums up the numbers from
the Junit output.
Maybe we're running some tests twice as part of a suite and Jenkins merges
those results?
> Increase test case stability in pre-commit job
> ----------------------------------------------
>
> Key: OOZIE-2817
> URL: https://issues.apache.org/jira/browse/OOZIE-2817
> Project: Oozie
> Issue Type: Bug
> Reporter: Peter Cseh
> Assignee: Peter Cseh
> Fix For: 5.0.0
>
> Attachments: OOZIE-2817-001.patch, OOZIE-2817-002.patch,
> OOZIE-2817-003.patch, OOZIE-2817-amend.patch
>
>
> We could collect the failed tests and rerun them in the pre-commit hook.
> They will still fail if they are broken, but they'll probably pass if they
> are just flaky.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
